Constant time median filtering of extra large images using Hadoop Mohammadreza Azodinia, Vahid Farrokhi, Andras Hajdu University of Debrecen m.r.azodinia@inf.unideb.hu vahid.farrokhi@inf.unideb.hu hajdu.andras@inf.unideb.hu Abstract The spatial resolution of remote sensing and medical images such as MRI, CT and PET are constantly increasing and analyzing these images in real time is a challenging task. But this limits the efficiency of many image processing algorithms. Among different efficient image processing algorithms, median filtering is a principal element in many image processing situations which manages to reduce the noise while preserving the edges. Median Filtering in Constant Time (MFCT) is a simple yet fastest median filtering algorithm which can handle N-dimensional data in fields like medical imaging and as- tronomy. With trend toward the median filtering of large images and pro- portionally large kernels, Hadoop MapReduce (a popular big data processing engine) can be applied and utilized. MapReduce provides the simplicity of defining the map and reduce functions while the framework takes care of parallelization and failover automatically. Hence, in this paper we discuss on possibility of the incorporation of MFCT algorithm with Hadoop MapReduce framework to improve the per- formance of processing of extra large images. Keywords: median filtering, MFCT, MapReduce, Hadoop, parallelization MSC: 68-06, 68U10 1. Introduction The daily growth of the amount and size of data generated by modern acquisition devices is in a challenge with the processing capabilities of single devices. Consid- ering the fact that processor architectures are reaching their physical limitations, using distributed computing technologies would be the best and safest choice for solving those problems that do not fit into the capability of a single machine. There- fore, relating large-scale data processing, as a solution many approaches turned Proceedings of the 9 th International Conference on Applied Informatics Eger, Hungary, January 29–February 1, 2014. Vol. 1. pp. 93–101 doi: 10.14794/ICAI.9.2014.1.93 93